Mark ở vũ trụ khác · Another Universe Papers That Changed How We Learn #3 · Memory · Neurology

Patient H.M.

Scoville & Milner · Journal of Neurology, Neurosurgery and Psychiatry · 1957

The paper"Loss of recent memory after bilateral hippocampal lesions"
Findinglose the hippocampus on both sides and you cannot form new memories, yet intelligence stays
The patientH.M. (Henry Molaison), operated on in 1953 at age 27
LabelSolid, with two corrections

Source: Scoville and Milner (1957), J Neurol Neurosurg Psychiatry 20(1): 11–21; name, birth year and age at surgery from reviews of H.M. (Corkin 2002; Squire 2009).

Recap in 1 minute

Essence

One operation, 1953To treat severe epilepsy, a surgeon removed the inner temporal lobe on both sides of H.M.'s brain, including most of the hippocampus.
New events would not stayHis memory quotient fell to 67, where 100 is average. Whatever just happened slipped away once he was distracted.
Intelligence intactIQ 112 in 1955, higher than 104 before surgery. Remembering is not the same as thinking.
Skills still learnedTracing a star in a mirror, he improved over three days without remembering any practice.
Memory comes in kindsRemembering events needs the hippocampus. Remembering how to do things does not. That changes how we study.

Source: Scoville and Milner (1957) (surgery, IQ 104 and 112, memory quotient 67); Milner (1962), via the reviews of Squire (2009) and Corkin (2002) (mirror drawing).

Part 1

The paper's question

Where in the brain does memory live? Before 1957 many thought it was spread across the whole cortex. One operation that went tragically wrong pointed to a specific place, and showed that "remembering" is not one job but several.

The question and its framing are Mark's, based on the review by Squire (2009), Neuron 61: 6–9.

1953SurgeryBoth sides removed to treat epilepsy. Seizures decrease.
→
1955–1957MeasurementBrenda Milner tests him: memory broken, intelligence intact.
→
1962 to nowA map of memorySkills can still be learned. Memory has several systems.
Part 2

The story of the discovery

Who, where, how, against the medicine of the 1950s.

1926 · Henry Molaison is bornIn Connecticut, USA. Minor seizures began around age 10, major ones from 16. The drugs of the day could not stop them.
1953 · Surgery in HartfordNeurosurgeon William Scoville removed the inner temporal lobe on both sides, including most of the hippocampus and the amygdala. Henry was 27. Scoville himself called it a "frankly experimental" operation.
After surgery · Something no one expectedThe seizures eased. But H.M. could not find the way to the hospital bathroom, did not recognise people he had just met, and reread old magazines as if new.
April 1955 · Brenda Milner comes to testA psychologist from the Montreal Neurological Institute (Canada). IQ 112, higher than 104 before surgery. Memory quotient 67. He still thought the year was 1953.
1957 · The paperScoville and Milner compared H.M. with 9 other people given similar operations for mental illness. Everyone whose hippocampus was removed on both sides lost memory; the more removed, the worse.
1962 · The star in the mirrorMilner had H.M. trace the outline of a star while watching only its mirror image. Over three days he got better and better, each time saying he had never done it.
2008 · His real name is releasedHenry Molaison died on 2 December 2008, aged 82. Nearly 100 researchers had worked with him over more than 50 years.

Source: Scoville and Milner (1957) (surgery, 10 patients, IQ, memory quotient, April 1955; the "still thought it was 1953" detail per reviews of the case); Dossani et al. (2015), World Neurosurgery ("frankly experimental"); Milner (1962) via Squire (2009); Corkin (2002), Nature Reviews Neuroscience 3: 153–160 (nearly 100 researchers); birth year, onset ages and date of death per Wikipedia and Science Friday (secondary source). The bathroom and magazine details are as described in reviews of the case (secondary source).

Context in the 1950s: why such an operation?

It was the era of brain surgery for mental illness (such as frontal lobotomy). Scoville had operated on the inner temporal lobe of many patients with schizophrenia, with mixed results. For H.M., surgery seemed worth trying because epilepsy made a normal life almost impossible.

No one yet knew the hippocampus was needed for memory. The 1957 paper itself warned surgeons: do not remove the hippocampus on both sides.

Source: summary of the 2000 reprint in the Journal of Neuropsychiatry and Clinical Neurosciences 12(1); Dossani et al. (2015).

Part 3

What the paper says, in 3 sentences

1. Removing the hippocampus and nearby tissue on both sides of the brain takes away the ability to lay down lasting memories of new events.

2. Intelligence, language, immediate recall and most old memories remain.

3. Removing one side, or less tissue, causes little or no memory loss: severity follows the extent of removal.

H.M.'s scores. Intelligence: 104 before surgery (August 1953), 112 after (April 1955). Memory quotient after surgery: 67, where 100 is the scale's average. H.M.'s scores (100 = the scale's average) 100 IQ before (Aug 1953)104 IQ after (Apr 1955)112 memory quotient (1955)67 intelligence did not drop; memory fell far below average
Remembering and thinking are different jobs. Before H.M., memory loss usually came with general intellectual decline. In H.M. the two came apart. The paper says his IQ after surgery "compares favourably" with the score before.Source: Scoville and Milner (1957): IQ 104 measured by Dr Liselotte Fischer in August 1953; IQ 112 and memory quotient 67 measured by Milner in April 1955 (Wechsler scales). The memory quotient scale sets 100 as average.
Two memory systems. Left: memory for events and facts, needs the hippocampus, broken in H.M. Right: memory for skills such as mirror drawing, does not need the hippocampus, H.M. still learned over three days. The rising line is an illustration, not data. memory for events, facts "who I met yesterday", "what this word means" needs the hippocampus to record in H.M.: brokennew events do not stay memory for how, skills mirror drawing, cycling, typing relies on other regions in H.M.: still learns day 1 → day 3 (illustration)
The biggest finding came five years after the paper. In 1962 Milner showed H.M. learned a new skill normally, though he had no memory of practising. From then on scientists spoke of "multiple memory systems".Source: Milner (1962), Physiologie de l'hippocampe (CNRS), via the reviews of Squire (2009) and Squire and Wixted (2011), Annual Review of Neuroscience 34: 259–288. The rising line is an illustration, not measured data.
Part 4

How this paper is rated

Every row has a source. Citation counts say who counted and when.

PublishedJournal of Neurology, Neurosurgery and Psychiatry, 1957. Reprinted in 2000 in the "Neuropsychiatry Classics" series of the Journal of Neuropsychiatry and Clinical Neurosciences.Source: PubMed 13406589; J Neuropsychiatry Clin Neurosci 12(1): 103–113.
Citations5,284, per the "total cites" figure printed with Milner and Klein's 2016 retrospective in the same journal (no count date given). In 2009 Squire gave "nearly 2,500". Google Scholar: not checked (blocked).Source: Milner and Klein (2016), J Neurol Neurosurg Psychiatry 87(3): 230–231, PDF on the McGill site; Squire (2009), Neuron 61: 6–9. Looked up via web search on 15 Oct 2026.
Replicated?Not replicable as an experiment (nobody would repeat the surgery). But within the paper 10 patients showed the same pattern, and later cases of hippocampal damage and animal studies confirmed the hippocampus's role.Source: Scoville and Milner (1957); Squire and Wixted (2011).
Scientific critiqueMRI in 1997: the removal was about 5.4 cm (left) and 5.1 cm (right) long, not the 8 cm Scoville estimated; about 2 cm of posterior hippocampus remained. Post-mortem study (2014): an extra small lesion in the left frontal cortex and white-matter damage, so not every deficit can be pinned on the hippocampus alone.Source: Corkin et al. (1997), Journal of Neuroscience 17(10): 3964–3979; Annese et al. (2014), Nature Communications 5: 3122; Howard Eichenbaum quoted in Science News (secondary source).
Ethical critiqueAn experimental operation on a young man. In 2016 journalist Luke Dittrich (Scoville's grandson) published a book criticising how H.M. was studied and his records handled. MIT rejected this; more than 200 scientists signed a letter supporting researcher Suzanne Corkin. It remains a dispute, with no ruling by any body.Source: STAT 9 Aug 2016; MIT McGovern Institute; Undark 4 Oct 2016 (secondary source).
AwardsBrenda Milner received the Balzan Prize (2009) and the Kavli Prize in Neuroscience (2014, with John O'Keefe and Marcus Raichle). The H.M. case appears in nearly every psychology and neuroscience textbook.Source: McGill Neuro awards page for Brenda Milner; Squire (2009).
VerdictSOLID The main finding has held for almost 70 years. Two corrections: not all of the hippocampus was removed, and H.M.'s brain had other damage. Mark's assessment.
Part 5

Fact check: what is true and what is myth

14 things often said about H.M., checked against the original paper and later studies. Tap a row to see why.

Length of the removal. Scoville estimated about 8 cm. MRI in 1997 measured about 5.4 cm on the left and 5.1 cm on the right. About 2 cm of posterior hippocampus remained. Scoville's estimate (1957)8 cm MRI 1997, left side5.4 cm MRI 1997, right side5.1 cm about 2 cm of posterior hippocampus left (shrunken), not enough for normal memory
Measured by a surgeon's eye in 1953, by a scanner in 1997. The often-copied 8 cm is the estimate made during surgery.Source: Scoville and Milner (1957); Corkin et al. (1997), Journal of Neuroscience 17(10).
SOLIDAfter the 1953 surgery H.M. could barely form new memories of events

Memory quotient 67 in 1955; he did not recognise people met many times. Studies over the next 50 years agreed (Corkin 2002).

SOLIDHis intelligence did not decline

IQ 112 in April 1955 versus 104 in August 1953 before surgery (Scoville and Milner 1957).

SOLIDMemory loss tracked how much hippocampus was removed on both sides

Of the paper's 10 patients, only those with damage to the hippocampus and hippocampal gyrus on both sides lost memory; more removal, worse loss.

SOLIDH.M. still learned new skills without remembering the practice

Mirror-tracing a star, improving over three days (Milner 1962, via Squire 2009).

SOLIDImmediate recall (a few digits, a short sentence) was normal

The paper reports intact immediate memory; the loss showed after distraction or a few minutes (Squire 2009).

SOLIDThe surgeon himself called it experimental

Scoville used the phrase "frankly experimental" (quoted in Dossani et al. 2015).

CAUTION"The surgery removed 8 cm"

That was Scoville's estimate. MRI in 1997 measured about 5.4 cm and 5.1 cm.

CAUTION"The hippocampus stores all our memories"

The hippocampus is needed to record and consolidate new memories of events. H.M. still remembered his childhood, so old memories are not all stored there (Squire and Wixted 2011). Spens and Burgess (2024) model the hippocampus as a fast recorder that later "teaches" the cortex.

CAUTION"All of H.M.'s deficits came from losing the hippocampus"

The 2014 brain study found an extra small lesion in the left frontal cortex and white-matter damage (Annese et al. 2014).

CAUTION"His epilepsy was caused by a bicycle accident"

Often told, but a mild injury at age 7 is unlikely to be the cause; three family members had epilepsy (Tidsskriftet, Journal of the Norwegian Medical Association, 2026; secondary source).

CAUTION"A 2016 book proved misconduct in the H.M. research"

Those are Luke Dittrich's allegations. MIT rejected them and over 200 scientists signed a letter supporting Corkin. No official finding (STAT, Undark; secondary source).

FALSE"H.M. could never learn anything new"

He learned skills (Milner 1962) and even a little new knowledge: cued with first names, he gave the right surnames for 12 of 35 people who became famous after 1953 (O'Kane, Kensinger and Corkin 2004, Hippocampus 14: 417–425).

FALSE"His entire hippocampus was removed"

About 2 cm at the back remained, shrunken (Corkin et al. 1997; Annese et al. 2014).

FALSE"H.M. forgot his whole past"

He remembered his childhood; what he lost was mostly a short period before the surgery (Scoville and Milner 1957; Milner 1966, secondary source).

Source: as noted in each row; full list in Part 10.

Part 6

What the latest research says

From the 1957 paper to a computer model in 2024.

1997 · MRI scanThe removal was shorter than estimated; about 2 cm of posterior hippocampus remained but was not enough for normal memory.
2004 · He still learned a little knowledgeCued with first names, he gave the right surnames for 12/35 people famous after 1953, and picked famous names from fake ones about 87% of the time. Learning, but very slow and little.
2011 · Review "since H.M."Squire and Wixted sum up: there are several memory systems; the hippocampus and nearby areas are needed for events and facts, not for skills.
2014 · H.M.'s brain cut into 2,401 slicesJacopo Annese's team (San Diego) rebuilt the brain in 3D: hippocampal tissue remained, plus a small lesion in the left frontal cortex. The data are open to everyone.
2024 · The hippocampus as a "fast note-taker"Spens and Burgess (UCL) built a model: the hippocampus records each event quickly, then replays it so the cortex gradually learns general knowledge. It is a model, not direct evidence.

Source: Corkin et al. (1997), J Neurosci 17(10); O'Kane, Kensinger and Corkin (2004), Hippocampus 14(4): 417–425; Squire and Wixted (2011), Annu Rev Neurosci 34: 259–288; Annese et al. (2014), Nature Communications 5: 3122 (2,401 slices, cut in December 2009); Spens and Burgess (2024), Nature Human Behaviour, doi 10.1038/s41562-023-01799-z. No new 2023–2025 study on H.M.'s own data was found (searched 15 Oct 2026).

Part 7

Verdict

Scores are Mark's assessment based on the sources in Part 10, not data from a study.

Strength of the finding
5/5
Impact on science
5/5
Measurement accuracy (for 1957)
3/5
Directly useful to learners
3/5
Strengths
  • Standard scales, scores before and after surgery
  • 10 patients compared, pattern by extent of removal
  • Warned surgeons at once about the danger
Weaknesses
  • Removal estimated by eye, off by about 3 cm
  • One standout case, other damage hard to separate
  • Experimental surgery, many ethical questions

One man's tragedy gave a whole field its map: remembering events and remembering how travel by two different roads.

More: why does one case carry so much weight?

Usually a single case is weak evidence. H.M. is different for three reasons: exactly which areas were removed is known (and was later remeasured), there were scores from before surgery to compare, and he lived another 55 years, so many experiments could be run and rerun.

The paper was also not only about H.M.: 9 other people showed the same pattern. That is why the label is Solid, even though the 8 cm figure has been corrected.

The "More" box is Mark's commentary, based on sourced facts in Parts 2 to 6.

Part 8

Lessons: "knowing" and "being able to do" are two different kinds of learning

Your brain has at least two roads to memory. Send each thing you learn down the right one.

Four habits: separate what you need to know from what you need to do; practise skills by hand over many sessions; for knowledge, close the book and retell it; do not trust the feeling of fluency. 1 · two columnsknow / can doone list per subject 2 · skills: do itmany short sessionshands remember when heads forget 3 · knowledge: retellclose the book, say itretell again days later 4 · distrust fluencycan do ≠ understandask "why" at each step H.M. improved without knowing he had practised. You can use both roads.
Each habit rests on something the paper and later work showed. Skills are learned even when event memory fails (Milner 1962); knowledge of events must be recorded and consolidated (Squire and Wixted 2011).Source: findings in Parts 3 and 6; the habits are Mark's suggestions.
Two columnsWrite down what this subject asks you to "know" (concepts, formulas) and to "do" (solve, write, pronounce).
Skills: do it, in small dosesReading a worked solution does not replace solving. A little each day, many days, like H.M. and his star.
Knowledge: close the book, retellEvent memory has to be recorded. Retell without the book today, then again a few days later.
Distrust "it feels easy"Getting faster does not prove you understand. At each step ask "why this step?"
Self-check: the two-column test in 10 minutes

1. Pick a chapter you just studied. Draw two columns: "I can explain it" and "I can do it".

2. Left: close the book, write 3 key ideas in your own words. How many are right?

3. Right: solve a new exercise without the answer key. How far do you get?

4. Left good, right weak: you understand but have not practised. Right good, left weak: you are running on habit without knowing why.

5. Redo the test three days later. Review whichever column dropped more.

The test is Mark's suggestion, not a procedure from a study.

Part 9

Read next

Squire (2009), "The legacy of patient H.M. for neuroscience"Neuron 61: 6–9. Short and readable: why this case matters.
Suzanne Corkin (2013), Permanent Present TenseBy the researcher who worked with H.M. the longest. Read alongside Luke Dittrich's book (2016) for both sides.
Common Myths #9: "Memory works like a video camera"On Vũ Trụ Khác: why memory is rebuilt, not replayed. Read it
Papers That Changed How We Learn #04: Divers learning underwater (Godden and Baddeley 1975)Does where you learn affect what you can recall?
Part 10

Sources

The main sources are the original paper and peer-reviewed studies. The working environment blocked PubMed, journal sites and Google Scholar, so figures were checked against those sites' own excerpts in search results, without opening the full text. Secondary sources are marked.

  1. Scoville WB, Milner B (1957), "Loss of recent memory after bilateral hippocampal lesions", J Neurol Neurosurg Psychiatry 20(1): 11–21. PubMed · PDF (McGill) · reprint: J Neuropsychiatry Clin Neurosci 12(1): 103–113 (2000).
  2. Milner B, Klein D (2016), "Loss of recent memory after bilateral hippocampal lesions: memory and memories, looking back and looking forward", J Neurol Neurosurg Psychiatry 87(3): 230–231 (the 5,284 citations figure). PubMed
  3. Squire LR (2009), "The legacy of patient H.M. for neuroscience", Neuron 61(1): 6–9. ScienceDirect
  4. Corkin S (2002), "What's new with the amnesic patient H.M.?", Nature Reviews Neuroscience 3: 153–160.
  5. Corkin S, Amaral DG, González RG, Johnson KA, Hyman BT (1997), "H.M.'s medial temporal lobe lesion: findings from magnetic resonance imaging", Journal of Neuroscience 17(10): 3964–3979. jneurosci.org
  6. O'Kane G, Kensinger EA, Corkin S (2004), "Evidence for semantic learning in profound amnesia: an investigation with patient H.M.", Hippocampus 14(4): 417–425.
  7. Squire LR, Wixted JT (2011), "The cognitive neuroscience of human memory since H.M.", Annual Review of Neuroscience 34: 259–288.
  8. Annese J et al. (2014), "Postmortem examination of patient H.M.'s brain based on histological sectioning and digital 3D reconstruction", Nature Communications 5: 3122. nature.com
  9. Spens E, Burgess N (2024), "A generative model of memory construction and consolidation", Nature Human Behaviour. UCL Discovery
  10. Dossani RH et al. (2015), "The legacy of Henry Molaison (1926–2008) and the impact of his bilateral mesial temporal lobe surgery on the study of human memory", World Neurosurgery. ScienceDirect
  11. McGill Neuro, Brenda Milner awards page (Balzan 2009, Kavli 2014). mcgill.ca
  12. Secondary sources: Wikipedia "Henry Molaison" (birth year, onset); Science Friday (date of death); Tidsskriftet, "A life without a past" (2026) (bicycle accident, family history); Science News 2014 (Eichenbaum quote); STAT 9 Aug 2016, Undark 4 Oct 2016, MIT McGovern Institute (2016 dispute); UCL press release Jan 2024 (2024 model).